A Review of Computational Spinal Injury Biomechanics Research and Recommendations for Future Efforts by Reuben H. Kraft and Samantha L. Wozniak
ARL-TR-5673
Approved for public release; distribution is unlimited.
September 2011
NOTICES Disclaimers The findings in this report are not to be construed as an official Department of the Army position unless so designated by other authorized documents. Citation of manufacturer’s or trade names does not constitute an official endorsement or approval of the use thereof. Destroy this report when it is no longer needed. Do not return it to the originator.
Army Research Laboratory Aberdeen Proving Ground, MD 21005-5069 ARL-TR-5673
September 2011
A Review of Computational Spinal Injury Biomechanics Research and Recommendations for Future Efforts Reuben H. Kraft and Samantha L. Wozniak Weapons and Materials Research Directorate, ARL
Approved for public release; distribution is unlimited.
Form Approved OMB No. 0704-0188
REPORT DOCUMENTATION PAGE
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing the burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number.
PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS. 1. REPORT DATE (DD-MM-YYYY)
September 2011
2. REPORT TYPE
3. DATES COVERED (From - To)
I
Final
4. TITLE AND SUBTITLE
November 2010-April 2011 5a. CONTRACT NUMBER
A Review of Computational Spinal Injury Biomechanics Research and Recommendations for Future Efforts
W911QX-09-0057 5b. GRANT NUMBER 5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S)
5d. PROJECT NUMBER
Reuben H. Kraft and Samantha L. Wozniak
AH80 5e. TASK NUMBER 5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
8. PERFORMING ORGANIZATION REPORT NUMBER
U.S. Army Research Laboratory ATTN: RDRL-WMP-B Aberdeen Proving Ground, MD 21005-5066
ARL-TR-5673
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
10. SPONSOR/MONITOR'S ACRONYM(S)
11. SPONSOR/MONITOR'S REPORT NUMBER(S) 12. DISTRIBUTION/AVAILABILITY STATEMENT
Approved for public release; distribution is unlimited. 13. SUPPLEMENTARY NOTES
Author email:
[email protected] 14. ABSTRACT
Developing an understanding of the high rate injury biomechanics associated with blast-induced loading to the spine, lower extremities, central nervous system, and peripheral nervous system is critical to optimize protective equipment for Soldiers. Computational methods and high performance computing are an essential part of the toolset needed to develop the understanding of spinal injury caused by high strain rate loading. These methods provide a wealth of information that otherwise may not be available. While the research area of computational spinal injury biomechanics has received much attention in the past with regard to low strain rate degenerate effects, there are significant opportunities to enhance the state of knowledge of spinal injury in the area of high strain rate military loading. Through this literature review and careful contemplation, a list of six recommendations for future research directions has been developed.
15. SUBJECT TERMS
spinal injury, computational modeling, finite element analysis, explosion, blast 17. LIMITATION OF ABSTRACT
16. SECURITY CLASSIFICATION OF: a. REPORT
Unclassified
I
b. ABSTRACT
Unclassified
I
c. THIS PAGE
Unclassified
UU
18. NUMBER OF PAGES
48
19a. NAME OF RESPONSIBLE PERSON
Reuben H. Kraft 19b. TELEPHONE NUMBER (Include area code)
410-278-6142 Standard Form 298 (Rev. 8/98) Prescribed by ANSI Std. Z39.18
ii
Contents List of Figures
iv
List of Tables
v
Preface
vi
1. Introduction
1
1.1
Anatomy of the Spine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2
Osseous, Soft Tissue, and Neurological Injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2. Computational Spinal Injury Biomechanics
11
2.1
Models of the Vertebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2
Models of the Intervertebral Disc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3
Segment Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4
System Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5
Failure Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.6
Uncertainty Quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3. Conclusion and Recommendations
21
4. References
25
Appendix A. Verification, Sensitivity and Validation
35
Distribution List
36 iii
List of Figures
Figure 1. Bony anatomy of the (a) spinal column and (b) vertebra. Netter illustration from c www.netterimages.com. Elsevier Inc. All rights reserved. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Figure 2. Time scales associated with the loading process in the human body inside a vehicle due to mine detonation (17). Figure courtesy of the NATO Research and Technology Organization/Human Factors and Medicine Panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Figure 3. Principal applied loading mechanisms that may lead to spinal injury (21). Reprinted with permission from Springer Science + Business Media... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Figure 4. (a) Drawings show the grading of wedge vertebral fractures. The approximate degree of height reduction determines the assignment of grade to the fractured vertebra (reproduced with permission of Mary Ann Liebert, Inc.). (b) Lateral radiograph of lumbar spine shows a mild wedge fracture (grade 1) of L3 vertebra. (c) Lateral radiograph of lumbar spine shows moderate wedge fracture (grade 2) of L3 vertebra and moderate crush fracture (grade 2) of L2 vertebra. (d) Lateral radiograph of thoracic spine shows severe wedge fracture (grade 3) of T7 vertebra (31). Figures b through d reprinted with permission from the American Journal of Roentgenology. . . . . . . . . . . . . . . . . . . 9
Figure 5. (a) Patterns of the retropulsed fragment. (b) Sagittal and (c) axial CT scans of the thoracolumbar spine demonstrates an L4 burst fracture with retropulsion of bone into the spinal canal (as shown by arrows) (32, 33). Reprinted with permission from the American Journal of Roentgenology and Neurosurgical Focus. . . . . . . . . . . . . . . . . . . . . . . . . 10
Figure 6. Variations of Chance-type flexion-distraction spinal injuries include (a) classic Chance fracture, (b) fulcrum fracture, and (c) pure soft-tissue flexion-distraction injury. (d) Lateral radiograph shows a classic Chance fracture that extends horizontally through the pedicle (right arrow) and into the vertebral body. There is an anterosuperior wedge-shaped compression fracture (left arrow) in the same vertebral body (36). Reprinted with permission granted from The Radiological Society of North America and author. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
iv
List of Tables
Table 1. Compilation of various constitutive models and parameters found in recent computational models of the spine.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
v
Preface
Since this is a joint publication between a government employee and a government contractor, a break down of the report is offered to show that it is in the best interest of the government to have a combined government/contractor report. This review article was designed by Reuben Kraft to offer a precise overview of a subtopic within the large research field of spinal injury biomechanics, namely computational spinal injury biomechanics. Samantha Wozniak, a Bowhead, Inc. contractor, significantly contributed to the authorship as she gathered references and wrote summaries for targeted research topics dictated by Kraft. Kraft and Wozniak worked closely to shape this report each making many modifications of their own. If asked to divide the report into ”who did what,” it would be extremely difficult but would look close to the following: • Introduction: Kraft and Wozniak • Anatomy Section: Wozniak • Injury Section: Kraft and Wozniak • Vertebra Models Section: Wozniak • Intervertebral Disc Models Section: Wozniak • Segment Models Section: Wozniak • System Models Section: Kraft • Failure Models Section: Kraft • Uncertainty Quantification Section: Kraft • Conclusion and Recommendations Section: Kraft
vi
1.
Introduction
The incidence of spine injuries among American military personnel is currently estimated to be at its highest in comparison to the last 60 years (1). While injuries to the neck and spine only account for 5% of the total injuries seen in the current conflicts, these injuries are commonly the most debilitating and costly for the warfighter (2). Spinal injuries that result in spinal cord impingement are of particular concern because they may result in paralysis, loss of bowel and bladder function, inability to regulate automatic functions of the body, and many other complications. The classifications of spinal injuries are related to the abnormal forces acting on the central axial spinal pillar and the accompanying exaggerated movements that cause spinal instability. The magnitude and type of injuries produced by an explosion are a result of many factors including the type and amount of explosive material, delivery method, the surrounding environment, and the distance between the victim and the blast. Injuries induced by explosions can be divided into four classes: primary, secondary, tertiary, and quaternary blast injury. Primary blast injury is a result of the sudden increase in air pressure after the blast while secondary blast injury is due to projectiles from the blast. Tertiary blast injury is due to the thrust of the victim against stationary objects or wind disruption, and quaternary blast injury is a result of fire and heat generated by the explosion, toxic effects, or post-incident risks (3, 4). A study published in 2010 characterized combat-related spinal injuries sustained by a Brigade Combat Team (BCT), the basic deployable unit in the U.S. Army (1). The BCT involved in this study participated in The Iraq War Troop Surge of 2007, a major operation of the Iraq conflict. Twenty-nine of the 4,122 Soldiers deployed for 15 months sustained combat spine injuries (1). Twenty-three (74%) of these injuries were from improvised explosive devices. Four spine injuries were due to vehicular accidents and a gunshot wound led to one spinal injury. Specific spinal injuries sustained by these service members include lumbar burst fractures, compression fractures, nondisplaced type III fracture of the odontoid, traverse process fractures, and a flexion-distraction injury. Seven Soldiers were considered polytraumas, including one who experienced neurologic compromise due to a burst fracture.
1
The purpose of this report is to review recent literature on computational spinal injury biomechanics modeling for high strain rate loading, identify gaps in the research, and formulate recommendations for future efforts. This review does not aim to be an exhaustive list of papers that use finite computational techniques to analyze the spine, but rather focuses on current state-of-the-art techniques with regard to high rate injury and failure modeling. Models of the spine range from the sub-micron scale of vertebral bone to models of the whole spine and surrounding structures. This review focuses on models of individual spinal components and small segments of the spine at the macro scale. First, in sections 1.1 and 1.2, we present the spinal anatomy and a biomechanical description of spinal injuries sustained by Soldiers during combat, including the length and time scales associated with blast-induced spinal injury. Next, in section 2, a brief history of spine modeling is discussed, followed by a more in depth review of vertebra, intervertebral disc, segment, and full spine finite element models. In addition, section 2 includes a discussion of computational approaches for modeling spinal injury mechanisms, such as fracture and various techniques for handling uncertainty quantification. Finally, in section 3 we conclude with a summary of the state-of-the-art finite element model techniques and suggest recommendations for future research. 1.1
Anatomy of the Spine
The spine has many functions including to protect the spinal cord and associated nerves, allow for movement, support the body frame in an upright position, allow for flexibility, provide a structural foundation for the shoulder and pelvic girdles, prevent and absorb shock, and provide a structural base for rib attachments which protect the heart and lungs. The spinal column is an intricate structure, comprised of vertebrae and cartilage discs, known as intervertebral discs, stacked alternatively on top of each other. The column starts at the base of the skull and continues to the pelvis. As seen in figure 1a, the spine is divided into five regions: the cervical spine, the thoracic spine, the lumbar spine, the sacrum, the coccyx. Each individual vertebra is named by referring to the first letter of the region (e.g., “L” for lumbar), and, starting with the most superior vertebra in that region, numbered consecutively until the most inferior vertebra in the region has been named.
2
Figure 1. Bony anatomy of the (a) spinal column and (b) vertebra. Netter illustration from c www.netterimages.com. Elsevier Inc. All rights reserved.
Vertebrae differ in size and shape between regions and also within a region. For example, the 12 vertebrae in the thoracic spine (commonly referred to as T1 -T12 ) each have two thoracic facet joint sets– one set on the top and one set on the bottom– which are responsible for providing the middle back with the flexibility and strength it needs. A typical vertebra consists of two main parts: an anterior vertebral body and posterior vertebral arch. As seen in figure 1b, the lumbar vertebral arch is formed by a pair of pedicles, and laminae and supports seven processes: four articular, two transverse, and one spinous (note that T1 -T12 have six articular processes). The principal biomechanical function of the vertebral body is to support the daily loads of the spine. Accordingly, the vertebral bodies graduate in size from the cervical to the lumbar region. The microstructure of the vertebral body is comprised of highly porous trabecular bone and a thin, dense cortical shell. The trabecular bone supports the majority of the vertical 3
compressive loads while the outer cortical shell helps resist torsion and shear forces (5). A thin layer of semi-porous subchondral bone forms a boundary layer between the cancellous (synonymous with trabecular bone) core of the vertebral body and the intervertebral disc, known as the vertebral end-plate. The end-plates of a healthy disc prevent the highly hydrated nucleus from bulging into the adjacent vertebral bone, while simultaneously absorbing hydrostatic pressure that results from mechanical loading of the spine (6). In many respects the structure is similar to that of articular cartilage, and while the end-plate cartilage is not connected directly to the bone of the vertebral bodies, it does have direct connections with the disc through the lamellae of the inner (medial) annulus fibrosus (7). In the adult disc the end-plates typically are less than 1 mm thick, and although this varies considerably across the width of any disc, there is a tendency for the end-plate to be thinnest in the central region adjacent to the nucleus pulposus (8). Intervertebral discs lie between adjacent vertebrae and form a cartilaginous joint that allows for motion between the vertebrae. These structures help absorb compression forces, support bending loads and resist rotation, tension, and shear (9). Discs are complex structures made up of an outer ring of fibrous cartilage (type I and II), also known as the annulus fibrosus, and an inner gelatinous core termed the nucleus pulposus. The nucleus pulposus contains randomly organized collagen fibers and radially organized elastin fibers (10). The annulus fibrosis is comprised of alternating layers of type I (outer annulus) and type II (inner annulus) collagen at 30◦ angles to the end-plate. This arrangement leads to an anisotropic and inhomogeneous material. In addition, the orientation of the collagen fibers makes the structure nonlinear (11). Another load-bearing junction between the vertebrae is the zygapophysial joint (zygapophyseal, or facet joint). This joint is formed between the superior articular process of one vertebrae and the inferior articular process of the vertebrae directly above it as schematically depicted in figure 1b. In each spinal segment there are two facet joints, which, in combination with the disc, allows for motion including flexion, extension, rotation, and lateral bending. These joints also may restrict movement depending on their orientation; they limit flexion in the thoracic spine and rotation in the lumbar spine (12). Similar to other joints in the body, each facet joint is surrounded by a capsule of connective tissue and covered by articular cartilage. Within the joint capsule, the joint is lined by a membrane known as the synovium. Synovial fluid surrounds the joint and is compromised of hyaluronic acid, lubricin, proteinases, and collagenases. The lubricin found in the synovial fluid allows the cartilage to move with very little friction. The coefficient of friction for synovial joints in humans has been estimated to 4
be 0.01 for static friction and 0.003-0.0057 for kinetic friction (13, 14). Excessive movement of the spinal components is limited by the spinal ligaments. Ligaments, in combination with muscles and tendons, help preserve the natural articulation of the spine and stabilize the components. There are two primary ligament systems in the spine: the intersegmental and intrasegmental systems (15). The intrasegmental system holds individual vertebrae together and is made up of the ligamentum flavum, interspinous ligament, intertransverse ligament, and facet capsule. The intersegmental system holds many vertebrae together and includes the anterior longitudinal ligament, posterior longitudinal ligament, and supraspinous ligament. Within the vertebral canal is the spinal cord, which is protected by three tissue layers, known as meninges. Beginning as a continuation of cranial dura mater, the outer protective membrane, spinal dura mater, extends to the level of the second sacral vertebra (16). Between the dura mater and the vertebrae is a space, known as the epidural space, filled with adipose tissue and a network of blood vessels. The middle layer, or arachnoid mater, is separated from the inner layer (pia mater) by the subarachnoid space which contains cerebrospinal fluid. 1.2
Osseous, Soft Tissue, and Neurological Injury
It is important to understand the length and time scales associated with blast-induced spinal injury. In vehicle mine blasts, tertiary blast effects tend to be the most common injury mechanism (4). Within 0.5 ms of the initiation of the explosion, a shock wave hits the bottom plate of the vehicle and causes a large peak in pressure. This pressure results in local acceleration and subsequent deformations of the plate, which apply significant axial loads to the Soldier and may result in spinal, pelvic, and lower limb fractures (17). In 2007, the North Atlantic Treaty Organization (NATO) published a technical report, Test Methodology for Protection of Vehicle Occupants against Anti-Vehicular Landmine Effects, which provided information about the process of vehicular mine blasts (17). As described in the report, as the blast wave reflects under the vehicle, a pressure force acts on the bottom of the vehicle and causes it to lift off the ground. The acceleration of the vehicle and the collisions that follow are a significant source of injury, especially if the Soldier is not properly restrained. As seen in figure 2 (reprinted from the NATO technical report), the body is subjected to both the local effects (caused by shock and high rate deformation of the vehicle) and global effects (vehicle motion over time) of the vehicle mine detonation process. Drop-down effects and subsequent events, such as
5
rollover, are not considered due to a high variability in boundary conditions.
Figure 2. Time scales associated with the loading process in the human body inside a vehicle due to mine detonation (17). Figure courtesy of the NATO Research and Technology Organization/Human Factors and Medicine Panel.
Spinal injury in Soldiers is often caused by high accelerative loading and may result from explosions and vehicular accidents. Mechanisms of spinal injury include axial deformation (compression or tension); torsion or axial rotation; segmental translation (shear); and simultaneous or successive combination of the previous three mechanisms (18). Principal applied loading mechanisms that may lead to spinal injury are shown in figure 3. The thoracolumbar junction (TLJ), T12 -L2 , is a common site of traumatic injuries, accounting for 30–60% of all spinal injuries (19). However, in current military operations, for spinal injury cases involving underbelly blast to a vehicle, the common site of burst fractures is mid-lumbar. This is unusual for spinal fractures which tend to happen at C7 -T1 and T12 -L1 . This difference may be due to rate of loading, posture, and armor load on the Solider (12, 20).
6
Figure 3. Principal applied loading mechanisms that may lead to spinal injury (21). Reprinted with permission from Springer Science + Business Media.
Trauma to the spine may result in vertebral fractures, degenerative disc disease, spinal cord contusions, lesions of connecting ligaments, and other disabilities. Some common disabilities include protrusions of cartilage of the intervertebral disc through the vertebral body endplate into the adjacent vertebra (Schmol’s nodes); degenerative osteoarthritis of joints between the spinal vertebrae and neural foraminae (spondylosis, narrowing of space between vertebrae); and anterior displacement of vertebra in relation to the vertebrae below (spondylolisthesis) (22). Among adults, the most common injuries are vertebral column fractures and fracture dislocations (23). Aside from visible injuries to the spinal cord, such as fractures and crushing, Soldiers can also experience damage to the nerve roots or myelinated fiber tracts that are responsible for carrying signals to and from the brain. Injuries to the neck (C1 -C7 ) tend to be more life-threatening since they may affect the phrenic nerve, a nerve originating from the cervical spine that is responsible for the movement of the diaphragm (17). There is also increasing evidence that the central nervous system may be damaged as a direct effect of primary blast (24, 25). Systemic acute gas embolism, a result of pulmonary disruption from the blast wave, may result in occluded blood vessels in the brain or spinal cord (24). The spine is capable of withstanding large amounts of pressure (spinal injuries typically do not develop below accelerations of 20–25 G) and the construction of the neural arch in the vertebrae provides significant resistance to forward shear (26, 27). However, as the lateral trunk velocity increases, the lateral shear forces increase proportionally and more intensely than other spinal forces (27). Compression forces also increase as velocity increases, and thus when the combination of lateral shear and compression forces are too great for the vertebra to withstand, fractures may occur in the pedicle, lamina, transverse, or spinous process; facet; or vertebral body (17). The severity of these 7
fractures will depend on the structure stability of the fracture, i.e., whether or not the bony and ligamentous integrity is compromised such that free movement can occur. Even with a stable fracture, the spine is still capable of carrying and distributing a person’s weight. Unstable fractures involve an abnormal motion of the fracture site, which could potentially impinge the spinal cord or its vascular supply, resulting in pain or loss of neurological function. Neurologic deficit is present in 10–30% of all spinal column injuries (23, 28). Typically, the degree of associated neurological injury is correlated with the extent of the vertebral column fracture and any accompanying dislocation or subluxation. Axial compressive loads commonly cause compression, burst, or chance fractures. A wedge compression fracture occurs when there is a combination of a flexion bending moment and compressive loading, crushing the anterior aspect of the vertebra and forming a wedge shape (figure 4). Three subtypes of wedge fractures can be defined based on the height reductions in the anterior, middle, and posterior dimensions of the vertebra (29). The anterior column includes the anterior longitudinal ligaments and anterior one third of the vertebral body. The middle column consists of the posterior longitudinal ligament and posterior two thirds of the vertebral body. The posterior column is comprised of the facets, spinous processes, pedicles, and posterior ligamentous complex (19, 30). In a grade 1 fracture, only the anterior column is crushed (stable fracture). In a grade 2 fracture, the anterior column is crushed and the posterior column ligaments fail (potentially unstable fracture). The loss of anterior vertebral body height is typically 50%. A lateral radiograph of a grade 3 wedge fracture of the T7 vertebra can be seen in figure 4d.
8
(a)
(b)
(c)
(d)
Figure 4. (a) Drawings show the grading of wedge vertebral fractures. The approximate degree of height reduction determines the assignment of grade to the fractured vertebra (reproduced with permission of Mary Ann Liebert, Inc.). (b) Lateral radiograph of lumbar spine shows a mild wedge fracture (grade 1) of L3 vertebra. (c) Lateral radiograph of lumbar spine shows moderate wedge fracture (grade 2) of L3 vertebra and moderate crush fracture (grade 2) of L2 vertebra. (d) Lateral radiograph of thoracic spine shows severe wedge fracture (grade 3) of T7 vertebra (31). Figures b through d reprinted with permission from the American Journal of Roentgenology.
Burst fractures are typically associated with high energy trauma and are most commonly found between T5 and T8 and the TLJ (29). However, as mentioned earlier, in current military operations, for spinal injury cases involving underbelly blast to a vehicle, the common site of burst fractures is mid-lumbar. They are characterized by the crushing of the anterior and middle aspects of the vertebral body (figure 5). The state of the posterior elements determines whether the fracture is stable or unstable. Displacement of posterior elements and/or vertebral body or facet dislocation or subluxation define an unstable burst fracture (29). Sagittal and axial computed tomography (CT) scans of the thoracolumbar spine, seen in figures 5b and 5c, demonstrates an L4 burst fracture with retropulsion of bone into the spinal cord. Displacement of bone fragments may lead to life-threatening injuries if they penetrate into the spinal canal and cause loss of spinal cord function (17). While burst fractures only account for 15% of spinal injuries, patients with burst fractures can reach up to 50–60% neurological deficit from their pre-injured state (19).
9
(a)
(b)
(c)
Figure 5. (a) Patterns of the retropulsed fragment. (b) Sagittal and (c) axial CT scans of the thoracolumbar spine demonstrates an L4 burst fracture with retropulsion of bone into the spinal canal (as shown by arrows) (32, 33). Reprinted with permission from the American Journal of Roentgenology and Neurosurgical Focus.
While experimental studies have identified sites of failure initiation in burst loading within the literature, there is a lack of consensus to which region fails first. Several authors believe that the initial fracture occurs in the vertebral body, which leads to expulsion of bone and fluid into the spinal canal. This results in reduced support of the endplates, continuing their exposure to compression and eventually forcing them to crack (34). Others report that an increase in pressure within the nucleus pulposus causes the endplates to bulge and crack (19). Hongo et al. found high compressive and tensile strains at the base of the pedicle of T10 , L1 , and L4 , indicating that the base of the pedicle is the site of burst fracture initiation (35). Flexion-distraction thoracolumbar spine fractures, also known as Chance fractures, commonly occur at the thoracolumbar junction and tend to be the result of violent forward flexion (bending), which causes a distraction injury (pulling apart) to the posterior elements of the vertebra. Chance fractures occur when the anterior portion of the vertebral body is minimally compressed and there is a transverse fracture through the posterior elements of the vertebra and posterior portion of the vertebral body (2). Several subtypes of flexion-distraction thoracolumbar spine fractures exist, including a classic Chance fracture, fulcrum fracture, and pure soft-tissue flexion-distraction injury as seen in figures 6a, 6b, and 6c, respectively. The pathophysiology of each fracture depends on the axis of flexion.
10
(a)
(b)
(c)
(d)
Figure 6. Variations of Chance-type flexion-distraction spinal injuries include (a) classic Chance fracture, (b) fulcrum fracture, and (c) pure soft-tissue flexion-distraction injury. (d) Lateral radiograph shows a classic Chance fracture that extends horizontally through the pedicle (right arrow) and into the vertebral body. There is an anterosuperior wedge-shaped compression fracture (left arrow) in the same vertebral body (36). Reprinted with permission granted from The Radiological Society of North America and author.
2.
Computational Spinal Injury Biomechanics
Biomechanical models of the human body have the aptitude to support advancements in many fields including medicine, biology, and engineering. Over the past 60 years, a substantial effort has been dedicated to modeling the human body. An assortment of models that vary in fidelity have been developed and range from having one degree of freedom to models with millions of degrees of freedom. The spine is a complex component of the human skeleton, consisting of many intricate components with diverse material properties. Modeling of the spine provides opportunities to gain more understanding and a detailed description of temporal and spatial deformation processes, including failure of the intervertebral discs and vertebrae. In general, research groups started with relatively simple models of individual components of the spine and eventually developed models that represented larger sections of the spine and surrounding anatomy. Latham (1957) is considered the first to have developed a mathematical model of the spine; focused on studying pilot ejection, he used a one-degree-of-freedom model consisting of two masses, representing the body and ejection seat, interconnected by a spring to represent the response of the spine to vertical acceleration (37–40). In Payne’s (1961) subsequent model, the head and upper 11
torso were modeled as a rigid body and the spine was represented by a spring-dashpot in a parallel system (38). Liu and Murray (1966) contributed to the modeling approach by proposing a mathematical model that consisted of a uniform rod and a rigid end-mass, representing the spinal column and head, respectively (41). It wasn’t until the late 1960s that segmentation of vertebrae and intervertebral discs were incorporated into models. Toth (1966) developed an eight-degree-of-freedom model in which vertebrae were modeled as rigid masses interconnected by springs and dampers, simulating the action of discs (42). A mathematical model developed by Orne and Liu (1971) took into account the curvature of the spine as well as inertial properties of the vertebral bodies and axial shear and bending stiffness were introduced (43). The interaction between the spine and the torso, however, were not considered. Prasad and King (1974) further developed this model to include the interactions of the articular facets (44). Up until this point, models did not take the rib cage or viscera into account. Belytschko et al. (1976) developed the first three-dimensional representation of the entire spine and several other models with increasing complexity (45). These finite element models consisted of the head, vertebrae (lumbar, thoracic, and cervical), ribs from T1 to T10, intervertebral discs, articular facets, ligaments, muscle, connective tissue, upper and lower viscera, and the pelvis. Skeletal components were modeled as rigid masses, intervertebral discs as beam elements, ligaments as spring elements, and articular facet interactions as hydrodynamic elements (38, 40, 46). Finite element modeling continues to be popular in spinal research. When constructing a finite element model, three primary areas need to be addressed: geometric representation including segmentation of spinal components; constitutive models and corresponding material properties of the components; and boundary conditions used to apply loading and constraints. More recently, model components are developed with geometries that match that of a specific in vitro specimen or that represent an average vertebra. It is important to define the geometry of the structure as accurately as possible because it greatly influences deformation (47–49). Generic models are developed from anatomical measurements while specific geometries can be formed from CT and magnetic resonance imaging (MRI) scans. Software packages convert data obtained from these images into a format that can be interpreted by solid modeling software and then input into finite element software. Wilcox et al. (2007) found that finite element results demonstrated a greater agreement with previous experimental results when the 12
models were morphologically accurate as opposed to generic models (47). It should be noted, however, that the accuracy of image-based models is affected by the original resolution of the medical image (50). Constitutive models and their parameters are also a critical aspect of developing a finite element model. Determining the properties of hard and soft tissues, and fluids, as well as discovering their complex interactions and changes over time, are a new set of challenges currently facing researchers. To date, a wide range of material values have been used in finite element models of the spine due to the large assortment of measurement techniques, testing strain rates, and sample preparation (table 1). The verification of mesh and boundary conditions and the sensitivity of results to material properties, however, has been limited (47) .
13
Table 1. Compilation of various constitutive models and parameters found in recent computational models of the spine.1 Anatomic Component
Material Model
Material Properties
References
Trabecular bone
Plastic with kinematic yield, Poroelastic
E = 0.1-2.2 GPa v = 0.2 σ Y = 42-110 MPa
(34, 51–56)
Cortical bone
Plastic with kinematic yield, Elastic
E = 5-22 GPa v = 0.3 σ Y = 132 MPa
(34, 52–55)
Posterior elements
Elastic
E = 1000-3500 MPa v = 0.25
(52–55, 57)
Cartilaginous Poroelastic endplates
E = 20-25 MPa v = 0.1-0.4
(34, 53, 54)
Facet cartilage
Elastic
E = 11 MPa v = 0.4
(53)
Nucleus pulposus
Poroelastic, Hyperelastic
K = 1.667 GPa E = 0.15-4 MPa v = 0.4999
(34, 52–55, 57, 58)
Annulus fibrosis
Orthotropic elastic, Hyperelastic, Poroelastic
E = 2-450 MPa v = 0.3-0.45
(51, 53–55, 58, 59)
Anterior longitudinal ligament
Hyperelastic
σ Y = 22.5 MPa εY = 0.2 E = 7.8-20 MPa v = 0.3
(54, 60, 61)
Posterior longitudinal ligament
Piecewise linear, Hyperelastic
σ Y = 16-20 MPa εY = 0.3-0.45 E = 10-20 MPa v = 0.3
(34, 52, 54, 60, 61)
E = 3.4-15 MPa v = 0.3
(54, 55, 61)
Vertebra
Intervertebral disc
Ligaments
Supraspinous Hyperelastic, ligament Nonlinear elastic
1
Spinal cord
Elastic
E = 0.26-1.3 MPa v = 0.35-0.49
(34, 62)
Dura mater
Anisotropic elastic
Err = Eθθ = 142 MPa Ezz = 0.7 MPa
(34)
Not all parameters are listed for each constitutive model. Furthermore, not all spinal components are listed. E is the Young’s Modulus, ν is Poisson’s ratio, σY is the yield strength, Y is the yield strain, and K is the bulk modulus. The symbols rr, θθ, and zz denote directions in the polar coordinate system.
14
2.1
Models of the Vertebra
The vertebral body have several features that should be considered in the development of its simulation. The vertebral body is comprised mainly of porous cancellous bone, with denser cortical bone at the periphery (11). Several groups have shown that varying material properties of the cancellous and cortical bone that make up the vertebral body, as well as changes to the material properties of the cartilaginous endplates, do not result in a significant change in biomechanical response (11, 51). However, these studies did not consider high-rate failure where fracture may be sensitive to inhomogeneous material properties. Thus, it may be important to consider in the future. Another important feature of vertebrae are their complex posterior geometry, which presents a challenge when meshing that area. Some researchers have simplified the geometry by developing models that exclude posterior elements and focus solely on the vertebral body (47, 58, 63–66). Computational models of isolated single vertebrae allow researchers to investigate the natural properties of the vertebra, without having to take into consideration soft tissue or ligament interactions. The principal clinical application for detailed modeling of a single vertebra is to provide insight into the loading regime within the spine. Current models are able to accurately predict compressive strength and are typically validated through direct validation by comparing model predictions with stiffness and strength values measured in vitro (see the appendix for a review on computational model verification, sensitivity and validation) (54). Single vertebrae models have also been developed to study the effects of metastatically compromised vertebral segments (58, 67) as well as simulate the presence of bone cement after a vertebroplasty procedure (65, 66, 68, 69). In the future, researchers would like to use these finite element models to predict fracture risk in vivo (54). The ability to non-invasively assess vertebral strength is critical for the diagnosis and management of osteoporosis. The current clinical standard to predict vertebral strength is to perform a bone mineral density (BMD) assessment through the use of dual energy x-ray absorptiometry (DXA) (70). Finite element models derived from quantitative computed tomography (QCT) scans, however, have been shown to better predict the vertebral bone strength than BMD techniques due to their ability to assess geometry and include heterogeneous mechanical properties of the bone (70–72). Several groups have loaded vertebral specimens at strain rates up to 0.5% until failure and then analyzed finite element models under matching conditions. The 15
models were able to predict yield loads, strain values, and stiffness values, which can be used to predict fracture sites. Model predictions were found to correlate with experimentally measured values and fracture sites (71–74). 2.2
Models of the Intervertebral Disc
Similar to vertebral simulations, one major reason for developing finite element models of the disc is to study the diseased state. Unlike vertebral modeling, which tends to focus on predicting fracture risk, intervertebral disc research has focused on understanding the process of degeneration and its biomechanical effects on the tissue (54). Some research groups have also investigated treatments for degenerative disc disease including replacement materials for the nucleus pulposus (75) and total disc replacement implants (76). Generally the geometry of intervertebral disc finite element meshes are simplified. Many researchers, for instance, assume the structure to be axisymmetric (11, 75, 77–81) or exhibit symmetry in the sagittal (80) or transverse planes (77). In addition, cranial and caudal surfaces of the disc are often assumed to have a flat surface although some groups are incorporating more realistic curvatures into their models (81). Recently, increasingly complexity has been incorporated into material models of the intervertebral disc tissue, including ways to represent fluid content and flow, osmotic forces, and anisotropy (77, 78). Validation of intervertebral disc model predictions often presents as a challenge due to the difficulties in taking in vivo measurements and preserving a realistic tissue environment in vitro. Single studies are often used for indirect validation, however, there tend to be large standard deviations in experimental data. It is often the case that reasonable agreement can be found but not quantified (54). 2.3
Segment Models
The motion segment, or functional spinal unit, is comprised of two adjacent vertebra, the interconnecting intervertebral disc, facet joint, and associated spinal ligaments. Since the spine may be considered a structure composed of several motion segments connected in series, motion segments are often used to study the biomechanical behavior of the spine. The functional spinal unit, along with larger spinal segments, can be used to simulate different spinal conditions and assist in the design and analysis of new spinal instrumentation, including pedicle screws, Harrington rods, and interbody cages. 16
In addition to modeling the vertebral and disc components, these models incorporate the facet joints and spinal ligaments and therefore present additional challenges beyond those seen in the previous two sections due to the limited experimental data available for characterization and the need to incorporate contact. Due to limited experimental data on the behavior of the facet joint, researchers have developed a number of approaches to model facet interactions. Some groups have incorporated material properties through the use of gap elements that adjust the force transferred across the joint based on the gap size (55, 82), while others included and assigned properties to a cartilage layer (53). The articular facets surfaces are generally assumed to be frictionless (53, 57, 83) or a low coefficient of friction has been applied (84). In general, researchers have depended on indirect validation to provide evidence of their model accuracy. A few authors, however, have directly compared their model’s results with tests on cadaveric specimens (51, 52, 55, 59, 85). All these studies, however, compare a generic model with experiments as opposed to a direct subject-specific segment model validation (54). Advanced models are now beginning to incorporate poroelastic characteristics to simulate the fluid flow between the intervertebral disc and cancellous core. This complex interaction is believed to play a major role in absorbing impact energy and may be partially responsible for a perceived rate-dependency of spinal component response. In a poroelastic model, the disc is modeled as having two distinct phases: a fluid phase and a permeable solid phase. Simon and Wu were among the first to include poroelastic behavior in a finite element segment model (86). From short- and long-term creep analyses, they found that failure is likely to be initiated in the endplates or in the cancellous bone adjacent to the endplate. Wu and Chen created a similar model that incorporated posterior elements and found that under long-term creep loading their model demonstrated similar results as Simon and Wu (87). A poroelastic finite element model of the L3 -L4 segment developed by Lee et al. (described in section 2.5) also supported Simon and Wu’s results (88). In 2007, Natarajan et al. developed a poroelastic model that incorporated physiological parameters, such as the change in permeability in the disc due to strain, in order to predict the failure initiation and progression in a lumbar disc due to cyclic loads. The model accurately predicted change in disc height during loading and unloading that was comparable to in vivo experiments (86). Many of the finite element models of the spine often assume a simplified geometry due to the complex nature of the vertebra and intervertebral disc. Facet joints and spinal 17
ligaments may also be disregarded even though these structures have an important influence on motion segment behavior (67). In addition, trunk muscles are often not included in finite element models even though they are known to be primary stabilizers in the spine (19). Muscle forces have a strong influence on spinal load and stresses in the intervertebral discs and, in their absence, the applied stress distribution in the discs differs greatly (85). This elimination and simplification of geometry may influence which area will fail first under dynamic loading, altering the mechanism of failure. It is therefore essential that future efforts include the complex geometry of the vertebra and intervertebral discs, as well as the ligaments and muscles. 2.4
System Models
System level models of the spine include lumbar, thoracic, and cervical spine sections, as well as coupling to other body regions that have an influence on the loads transmitted to the spine, such as the torso. Developing a high-resolution finite element model of the entire human spine and its connected anatomy is computationally expensive and is difficult to validate. In the past, full models of the spine were referred to as simple models because they lacked a high degree of anatomic fidelity– many use just a few (usually one or three) elements to represent each motion segment. Many of these efforts were reviewed in the beginning of section 2. Now, with the advances in high performance computing, it is becoming easier to incorporate more accurate geometry and improved material constitutive models. However, from the published literature, it seems that there are limited high-resolution full spine models that capture the effects of high strain rate loading regimes. Nevertheless, there are a few research groups making significant progress achieving this goal, for example the Southwest Research Institute (89) and the Spine Research Center at Walter Reed Medical Center (20). 2.5
Failure Models
An adequate comprehension of fracture mechanisms is essential for effective prevention, mitigation and treatment. Predicting vertebral fractures, however, continues to be a challenge due to the complex and elaborate geometry of the vertebra and the different material properties of the vertebral components. While a number of researchers have studied the instability, spinal canal impingement, and treatment methods of vertebral fractures produced by spinal trauma, there is still a lack of knowledge on the internal transient change of the vertebral bodies during fracture (19). Lee et al. used a three-dimensional, poroelastic finite element model of the L3 -L4 segment to predict
18
changes in biomechanical parameters such as dynamic stiffness, intradiscal pressure, stress in the endplates, and the shock-absorbing mechanism of the spine under different impact durations and loading rates (88). Additionally, the authors analyzed the relation between the loading rate and the potential of vertebral body fracture. It was found that short-duration loading has a strong effect of initiating fractures in the endplates or the posterior wall of the cortical shell. The authors also concluded that pressure in the nucleus is independent of the impact duration and depends on the magnitude of the impact force. To date, very little research has been performed to analyze the mechanism of fracture, especially on a predictive basis. Qiu et al. developed a finite element model of the thoracolumbar junction and subjected it to dynamic vertical impacts in order to investigate the vertebral burst fracture process (19). The authors found that an increase in pressure inside the nucleus forces each endplate to bulge towards the cancellous core of the vertebral body. When the endplates fail, the nucleus material is able to enter the vertebral body, pressurizing it more. This increase in pressure squeezes the contents of the vertebral body, including fat and marrow, out of the cancellous bone. When the nucleus material enters the vertebral body faster than the rate at which fat and marrow are being expulsed, a burst fracture is said to occur. Fracture sites predicted in this study were consistent with other studies as well as identical to the medical definition of burst fractures (19). Teo et al., developed a finite element model of the human atlas (C1) and were also able to accurately predict failure sites in comparison to experimental results (90). In 2009, El-Rich et al. developed a finite element model of the L2 -L3 spinal segment that analyzed the response of the bony structures and soft tissues to five degrees of sagittal rotation at different rates until fracture occurred. The model allowed for a detailed evaluation of failure occurrence and propagation over the bone. Results showed that the endplate, pedicle, and facet surface were the weakest regions when subjected to rapid movement in the sagittal plane and spinal injuries resulted from a sagittal rotational velocity exceeding 0.5◦ /ms (56). Segments subjected to the highest velocity resulted in a more severely damaged structural response compared to those subjected to the low and intermediate velocities, confirming that rapid movements increase the risk of injury (91). Under the highest velocity the model exhibited a considerable increase in intradiscal pressure in the nucleus and the stresses over the rest of the structures while there was only slight changes under lower and intermediate loading rates.
19
The spinal cord may be injured through various injury patterns; however, the relationship between spinal column injury pattern and spinal cord damage is not well understood. Greaves et al. developed a finite element model of the cervical spine and spinal cord in order to compare spinal cord distributions between three different injury mechanisms (62). Their model predictions showed that strain that induces distraction injuries were uniformly distributed lengthwise throughout the spinal cord, while dislocation injuries and contusion injuries (such as a burst fracture) led to focal strains. A biomechanical description of the injury process can help predict subsequent acute and chronic pathophysiology of spinal cord trauma. Li et al. developed a finite element model of the cervical spinal cord to simulate the complex injury mechanism of a hyperextension injury (92). The results of their hyperextension injury simulation showed high localized stress at the anterior and posterior horn in the gray matter of the spinal cord. It is believed that these stresses account for the predominance of hand weakness in patients with central cord injury (92). 2.6
Uncertainty Quantification
Probabilistic analysis methods provide a tool in which the biomechanical response due to characteristic uncertainties and variations in biological structures can be studied. The uncertainty of these values lead to an associated degree of uncertainty in the results (93). In order to improve the precision of computer models, probabilistic models, in which uncertainties and natural variations are incorporated into the model, have been developed. The fundamental concept of this method is that the input parameters are defined by an appropriate statistical distribution, not a single value (93, 94). A Normal or Gaussian distribution is commonly used in probabilistic models and defines the input variables by a mean and standard deviation. The values of each input parameter are sampled at random from the distribution and used in the model. The model can then be solved multiple times, allowing researchers to develop an indication of the range of the results that can be found. To date, this methodology has been used to study the influence of variations in material properties and geometry on spinal behavior (95–97). In 2005, the U.S. Naval Bio-dynamics Laboratory/U.S. Naval Air Warfare Center Aircraft Division used NESSUS, a computer software system that combines finite element methods with advanced reliability methods in order to model uncertainties in geometries, material properties, and other variables, to evaluate the risk of cervical spine injury from
20
maneuvers experienced during flight, high speed ejection, and crashes (97). The group incorporated mechanical and geometric data in order to develop, verify, and validate a parameterized probabilistic spinal injury prediction model. In order to simulate heterogeneous and nonlinear materials the group defined the mechanical characteristics of vertebrae, discs, ligaments and muscles through data collected from QCT, bone mineral density data, and cryomicrotomy (98). In addition, probabilistic models have been developed to study the impact spinal procedures, such as vertebroplasty and the implantation of spinal stabilization devices including artificial discs and pedicle screws, have on spinal functions (94, 99, 100). Progress in probabilistic biomechanics will depend upon the identification of injury modes, data collection, and the development of validated models (39). Additional research is necessary to integrate random loading schemes and random representations of geometric variations. As Thacker et al. (101) point out, uncertainty and error quantification play a key role in model verification and validation. Uncertainties associated with the model input parameters include material behavior, geometry, loads, initial conditions, and boundary conditions. Ideally, each input parameter would be varied according to statistically significant data obtained from experiments.
3.
Conclusion and Recommendations
Developing an understanding of the high rate injury biomechanics associated with blast-induced loading to the spine, lower extremities, central nervous system, and peripheral nervous system is critical to optimize protective equipment for Soldiers. Finite element analysis is an essential part of understanding the biomechanics of spinal injury. These models provide a wealth of information that otherwise may not be available. While the research area of computational spinal injury biomechanics has received much attention in the past with regard to low strain rate degenerative effects, there are significant opportunities to enhance the state of knowledge of spinal injury in the area of high strain rate military loading. Through this literature review and careful contemplation, a list of six recommendations for future research directions has been developed. Future research concerning the finite element model of the spine should focus on the following aspects:
21
• Develop high-resolution anatomic finite element models that capture the geometric complexity of the spine in a probabilistic manner. Models should be three-dimensional and include all anatomic components of the spine necessary to capture the physics of interest, namely, injury mechanisms such as fracture and ligament rupture. Models should be constructed directly from high-resolution medical data such as x-rays, CT scans and MRI that enable individual-to-individual variations of the geometric representation. Finite element studies should use a probabilistic approach to include the variation of geometry ranges that are typical during postural changes, movement, gravity, aging, and extrinsic loading. • Develop a multilevel approach to apply loading conditions to the spine. Currently the direct loads, their distributions, and strain rate that the spine experiences during a military underbelly blast event is unknown. There is detailed understanding about the explosion and the loading to the vehicle, as well as the spinal injury outcome observed from the field; however, little is known about the processes in between these endpoints. It is critical to understand the process of load transfer, injury initiation, and propagation. In addition to the complexity of considering the spine itself, it is suspected that the lower extremity reactions to underbelly blast and spinal injuries are coupled and cannot be considered isolated systems. Also, as discussed in section 2.4, a coupling between the spine and the torso exists. Thus, the coupled systems direct the research to face the challenging task of considering a wide range of length and times scales. While on one hand research must consider the large length scale (on the order of meters) for coupled systems that include the lower extremities, spine, and torso, we must also consider the smaller length scales (on the order of micrometers), such as the randomly organized collagen fibers and radially organized elastin fibers found in the nucleus pulposus, to understand the injury mechanisms. The military spinal injury research community would benefit from a set of high strain rate military loading paradigms for components (ligaments, processes), subassemblies (isolated vertebral bodies and intervertebral discs), assemblies (motion segments consisting of vertebral bodies, interconnecting ligaments, and intervertebral discs), and system level (whole spine) models. Loading paradigms should include rate, direction, and duration. Once again, finite element studies should use a probabilistic approach to include the variation of loading that are typical.
22
• Develop, implement, and validate constitutive models for individual spinal component materials at the appropriate strain rate conditions. As table 1 demonstrates, an abundance of literature sources describe a wide range of material properties of spinal structures. This variation in material property values has resulted in a substantial deviation of model predictions and results. Furthermore, results currently obtained from finite element models are often limited by the assumption of homogeneous, isotropic, and linear elastic material properties. While these assumptions simplify the analysis, they do not capture the realistic mechanical behavior of the spine (19, 84, 85, 92). Ultimately, models need to incorporate the anisotropic, nonlinear, heterogeneous, and porous material properties seen in the human spine. For example, we believe that a poroelastic model is essential to characterize changes of the annulus matrix from a solid to fluid phase during high-rate loading conditions. Recently, it has been shown that standard high strain rate material characterization techniques are not adequate for bones or soft tissues (102, 103). Additionally, standard gripping techniques for dynamic tension tests do not work for biological materials, and material inhomogeneity and anisotropy complicate the interpretation of measured stress and strain data. New dynamic test methods and data analysis techniques are needed to provide applicable high rate and failure data on spinal materials for constitutive model development and calibration. • Hierarchical verifications and validation. Computational models and results should be compared at multiple length scales. Thacker et al. (101) provide a useful hierarchical verification and validation scheme that decomposes the spine into components (ligaments, processes), subassemblies (isolated vertebral bodies and intervertebral discs), assemblies (motion segments consisting of vertebral bodies, interconnecting ligaments and intervertebral discs), and system level (whole spine) models. The process should start at smaller length scales such as the components and progress to the last stage of system level models. This approach requires simulations and experiments to be conducted at each level. • Use novel computational approaches for capturing high strain rate compressive failure. There are limited computational approaches to capture injury mechanisms associated with accelerative loading to the human spine. Current U.S. Army Research Laboratory (ARL) efforts are focused on using dynamic insertion of rate-independent cohesive elements to capture fracture mechanisms. The limitations of this approach are quickly being realized due to the large 23
fragmentation, finite element mesh dependency, and material interface compatibility experienced during high rate accelerative loading. In addition, computational approaches to model muscle rupture, tearing, and other failure mechanisms are neither common nor validated within the tools used today. One may think of cohesive elements as a traditional approach to modeling failure in the computational setting. However, given the large deformations that biological materials exhibit during accelerative loading and the complexity of the anatomic geometry, cohesive approaches will only work within a limited regime of deformation. Therefore, novel simulation capabilities should be pursued. The Material Point Method (MPM) and the Extended Finite Element Method (XFEM) are two methods that seem feasible. • Integrate functional outcome predictions coupled with mechanical damage. It is also important that computational models extend past mechanical descriptions to include functional measures. A multidisciplinary approach combining mechanics-based simulation and modeling with the clinical field to develop computational approaches for modeling structural and functional deficits that may arise from high strain rate spinal injuries is needed. The community should develop time-evolving predictive models to understand musculoskeletal stability, pain, and rehabilitation strategies.
At this time, the focus of numerous research groups is to develop a three-dimensional model under a full range of loading schemes, while simultaneously considering nonlinear material and geometry solutions. Current and past models are limited by their selection of material models and simplified geometries that are unrealistic in comparison to the human spine. With a full understanding of spinal injury mechanisms, design criteria based on human injury, such as vertebral body fracture tolerances, can be developed and implemented in order to develop spinal injury mitigation strategies for the warfighter.
24
4.
References
1.
Schoenfeld, A. J.; Goodman, G. P.; Belmont, P. J. Characterization of combat-related spinal injuries sustained by a U.S. Army Brigade Combat Team during Operation Iraqi Freedom. The Spine Journal 2010, In Press.
2.
Ragel, B. T.; Allred, C. D.; Brevard, S.; Davis, R. T.; Frank, E. H. Fractures of the Thoracolumbar Spine Sustained by Soldiers in Vehicle Attacked by Improvised Explosive Devices. Spine 2009, 34 (22), 2400–2405.
3.
Kluger, Y.; Peleg, K.; Daniel-Aharonson, L.; Mayo, A. The Special Injury Pattern in Terrorist Bombing. Journal of the American College of Surgeons 2004, 199 (6), 875–879.
4.
Ramasamy, A.; Hill, A. M.; Hepper, A. E.; Bull, A. M. J.; Clasper, J. C. Blast Mines: Physics, Injury Mechanisms and Vehicle Protection. Journal of the Royal Army Medical Corps 2009, 155 (4), 258–264.
5.
Ferguson, S. Biomechanics of the Spine. In Spinal disorders: fundamentals of diagnosis and treatment; Boos, N., Aebi, M., Eds.; Springer: Heidelberg, 2008.
6.
Moore, R. J. The Vertebral End-Plate: What Do We Know?. European Spine Journal 2000, 9 (2), 92–96.
7.
Hukins, D. W. L. Disc structure and function. In The Biology of the Intervertebral Disc; Ghosh, P., Ed.; CRC Press: Boca Raton, 1988.
8.
Roberts, S.; Menage, J.; Urban, J. Biochemical and structural properties of the cartilage end-plate and its relation to the intervertebral disc. Spine 1989, 14, 166–174.
9.
de Jager, M. Mathematical Modelling of the Human Cervical Spine: A Survey of the Literature; WFW-93.027; Eindhoven University of Technology 1993.
10. Urban, J. P.; Roberts, S. Degeneration of the intervertebral disc. Arthritis Research & Therapy 2003, 5 (3), 120–130. 11. Rao, A. A.; Dumas, G. A. Influence of material properties on the mechanical behaviour of the L5 - S1 intervertebral disc in compression: a nonlinear finite element study. Journal of Biomedical Engineering 1990, 13, 139–151. 25
12. Spink, R. J. Personal Communication, U.S. Army Research Laboratory, Survivability/Lethality Analysis Directorate (SLAD), 2011. 13. Elert, G. Coefficients of Friction of Human Joints. http://hypertextbook.com/ facts/2007/ConnieQiu.shtml, (accessed: 3 January 2011). 14. Elert, G. Friction. http://physics.info/friction/, (accessed: 3 January 2011). 15. Eidelson, S. G. Spinal Ligaments and Tendons. http://www.spineuniverse.com/ anatomy/spinal-ligaments-tendons, (accessed: 18 February 2011). 16. Watson, C.; Paxinos, G.; Kayalioglu, G. The Vertebral Column and Spinal Meniges. In The Spinal Cord; Press, A., Ed.; Elsevier Ltd.: San Diego, CA, 2009. 17. NATO HFM-090 Task Group 25. Test Methodology for Protection of Vehicle Occupants against Anti-Vehicular Landmine Effects, NATO Research and Technology Organization, 2007, RTO-TR-HFM-090. 18. Iencean, S. M. Classification of spinal injuries based on the essential traumatic spinal mechanisms. Spinal Cord 2003, 41, 385–396. 19. Qiu, T.-X.; Tan, K.-W.; Lee, V.-S.; Teo, E.-C. Investigation of thoracolumbar T12-L1 burst fracture mechanism using finite element method. Medical Engineering & Physics 2006, 28, 656–664. 20. Dmitriev, A. E. Personal Communication, Spine Research Center, Department of Orthopaedics and Rehabilitation, Walter Reed National Military Medical Center, 2011. 21. McElhaney, J. H.; Nightingale, R. W.; Winkelstein, B. A.; Chancey, V.; Myers, B. S. Biomechanical Aspects of Cervical Trauma. In Accidental Injury: Biomechanics and Prevention; Vol. 2; Nahum, A. M., Melvin, J. W., Eds.; Springer: The Netherlands, 2002; p 331. 22. Taneja, N. Spinal Disabilities in Military and Civil Aviators. Spine 2008, 33 (25), 2749–2753. 23. Macias, M. Y.; Wolfa, C. E. Biomechanical Considerations for Early Operative Interventions in Vertebral Column Fractures and Dislocations. In Neurosurgical Emergencies; Brandenburg, B., Ed.; Thieme Medical Publishers, Inc.: New York, 2008. 26
24. DePalma, R. G.; Burris, D. G.; Champion, H. R.; Hodgson, M. J. Current Concepts: Blast Injuries. The New England Journal of Medicine 2005, 352, 1335–1342. 25. Horrocks, C. L. Blast Injuries: Biophysics, Pathophysiology and Management Principles. Journal of the Royal Army Medical Corps 2001, 147 (1), 28–40. 26. Shananhan, D. F. Human Tolerance and Crash Survivability; RTO-EN-HFM-113; NATO Research and Technology Organization 2004. 27. Marras, W. S.; Granata, K. P. Spine Loading During Trunk Lateral Bending Motions. Journal of Biomechanics 1997, 30 (7), 697–703. 28. Leahy, M.; Rahm, M. Normal Synovial Fluid. http://emedicine.medscape.com/ article/1267029-overview, (accessed: 11 January 2011). 29. Parizel, P. M.; van der Zijden, T.; Gaudino, S.; Spaepen, M.; Voormolen, M. H. J.; Venstermans, C.; Belder, F. D.; van den Hauwe, L.; Goethem, J. V. Trauma of the spine and spinal cord: imaging strategies. European Spine Journal 2010, 19 (Supplement 1), S8–S17. 30. Academy, I. M. Spine Biomechanics. http://www.insightmedical.org/ index.php?option=com content&view=article&id= 52&Itemid=68, (accessed: 16 February 2011). 31. Lenchik, L.; Rogers, L. F.; Delmas, P. D.; Genant, H. K. Diagnosis of Osteoporotic Vertebral Fractures: Importance of Recognition and Description by Radiologists. AJR, American Journal of Roentgenology 2004, 183, 949–958. 32. Atlas, S. W.; Regenbogen, V.; Rogers, L. F.; Kim, K. S. The Radiographic Characterization of Burst Fractures of the Spine. American Journal of Roentgenology 1981, 136 (2), 369–375. 33. Harrop, J. S.; Hunt, G. E.; Vaccaro, A. R. Conus medullaris and cauda equina syndrome as a result of traumatic injuries: management principles. Neurosurgical Focus 2004, 16 (6), 19–23. 34. Wilcox, R. K.; Allen, D. J.; Hall, R. M.; Limb, D.; Barton, D. C.; Dickson, R. A. A dynamic investigation of the burst fracture process using a combined experimental and finite element approach. European Spine Journal 2004, 13, 481–488.
27
35. Hongo, M.; Abe, E.; Shimada, Y.; Murai, H.; Ishikawa, N.; Sato, K. Surface Strain Distribution on Thoracic and Lumbar Vertebrae Under Axial Compression: The Role in Burst Fractures. Spine 1999, 24 (12), 1197–1202. 36. Groves, C. J.; Cassar-Pullicino, V. N.; Tins, B. J.; Tyrrell, P. N. M.; McCall, I. W. Chance-Type Flexion-Distraction Injuries in the Thoracolumbar Spine: MR Imaging Characteristics. Radiology 2005, 236, 601–608. 37. Latham, F. A study in body ballistics: seat ejection. In Proceedings of the Royal Society of London. Series B, Biological Sciences; Vol. 147; Harrison and Sons: London, 1957. 38. Dimnet, J. Biomechanical Models of the Head-Neck System. In The Head-neck sensory motor system; Berthoz, A., Graf, W., Vidal, P. P., Eds.; Oxford University Press, Inc.: New York, 1992. 39. Thacker, B. H.; Kumaresan, D. P.; Kumaresan, S.; Yoganandan, N.; Pintar, F. A. Probabilistic finite element analysis of the human lower cervical spine. Mathematical Modelling and Scientific Computing 2001, 13, 12–21. 40. Fagan, M. J.; Julian, S.; Mohsen, A. M. Finite element analysis in spine research. Proceedings of the Institution of Mechanical Engineers, Part H: Journal of Engineering in Medicine 2002, 216 (5), 281–298. 41. Liu, Y. K. Towards a Stress Criterion of Injury- An Example in Caudocephalad Acceleration. Journal of Biomechanics 1969, 2, 145–149. 42. Toth, R. Multiplying degree of freedom, non-linear spinal model. In Proceedings of the 19th Annual Conference on Engineering in Medicine and Biology; 1966. 43. Orne, D.; Liu, Y. K. A Mathematical Model of Spinal Response to Impact. Journal of Biomechanics 1971, 4, 49–71. 44. Prasad, P.; King, A. I. An Experimentally Validated Dynamic Model of the Spine. Journal of Applied Mechanics 1974, 41, 546–550. 45. Belytschko, T.; Schwer, L.; Privitzer, E. Theory and Application of a Three-Dimensional Model of the Human Spine. Aviation, Space and Environmental Medicine 1978, 49, 158–165.
28
46. Williams, J. L.; Belytschko, T. B. A Three-Dimensional Model of the Human Cervical Spine for Impact Simulation. Journal of Biomechanical Engineering 1983, 105, 321–331. 47. Wilcox, R. K. The influence of material property and morphological parameters on specimen-specific finite element models of porcine vertebral bodies. Journal of Biomechanics 2007, 40 (3), 669–673. 48. Lu, M. Y.; Hutton, W. C.; Gharpuary, V. M. Can Variations in Intervertebral Disc Height Affect the Mechanical Function of the Disc?. Spine 1996, 21 (19), 2208–2216. 49. Maurel, N.; Lavaste, F.; Skalli, W. A Three-Dimensional Parameterized Finite Element Model of the Lower Cervical Spine. Study of the Influence of the Posterior Articular Facets. Journal of Biomechanics 1997, 30 (9), 921–931. 50. Yeni, Y. N.; Christopherson, G. T.; Dong, X. N.; Kim, D.-G.; Fyhrie, D. P. Effect of Microcomputed Tomography Voxel Size on the Finite Element Model Accuracy for Human Cancellous Bone. Journal of Biomechanical Engineering 2005, 127, 1–8. 51. Kumaresan, S.; Yoganandan, N.; Pintar, F. A. Finite element analysis of the cervical spine: a material property sensitivity study. Clinical Biomechanics 1999, 14, 41–53. 52. Guan, Y.; Yoganandan, N.; Zhang, J.; Pintar, F. A.; Cusick, J. F.; Wolfla, C. E.; Maiman, D. J. Validation of a clinical finite element model of the human lumbosacral spine. Medical & Biological Engineering & Computing 2006, 44 (8), 633–641. 53. Williams, J. R.; Natarajan, R. N.; Andersson, G. B. Inclusion of regional poroelastic material properties better predicts biomechanical behavior of lumbar discs subjected to dynamic loading. Journal of Biomechanics 2007, 40 (9), 1981–1987. 54. Jones, A. C.; Wilcox, R. K. Finite element analysis of the spine: Towards a framework of verification, validation and sensitivity analysis. Medical Engineering & Physics 2008, 30, 1287–1304. 55. Goel, V. K.; Grauer, J. N.; Patel, T. C.; Biyani, A.; Sairyo, K.; Vishnubhotla, S.; Matyas, A.; Cowgill, I.; Shaw, M.; Long, R.; Dick, D.; Panjabi, M. M.; Serhan, H. Effects of Charit´e Artificial Disc on the Implanted and Adjacent Spinal Segments Mechanics Using a Hybrid Testing Protocol. Spine 2005, 30 (24), 2755–2764. 29
56. El-Rich, M.; Arnoux, P.-J.; Wagnac, E.; Brunet, C.; Aubin, C.-E. Finite element investigation of the loading rate effect on the spinal load-sharing changes under impact conditions. Journal of Biomechanics 2009, 42, 1252–1262. 57. Schmidt, H.; Heuer, F.; Drumm, J.; Klezl, Z.; Claes, L.; Wilke, H.-J. Application of a calibration method provides more realistic results for a finite element model of a lumbar spinal segment. Clinical Biomechanics 2007, 22, 377–384. 58. Tschirhart, C. E.; Nagpurkar, A.; Whyne, C. M. Effects of tumor location, shape and surface serration on burst fracture risk in the metastatic spine. Journal of Biomechanics 2004, 37 (5), 653–660. ¨ 59. Pitzen, T.; Geisler, F. H.; Matthis, D.; Muller-Storz, H.; Pedersen, K.; Steudel, W.-I. The influence of cancellous bone density on load sharing in human lumbar spine: a comparison between an intact and a surgically altered motion segment. European Spine Journal 2001, 10 (1), 23–29. 60. Shim, V.; Liu, J. F.; Lee, V. S. A Technique for Dynamic Tensile Testing of Human Cervical Spine Ligaments. Experimental Mechanics 2006, 46, 77–89. 61. Kim, H.-J.; Chun, H.-J.; Kang, K.-T.; Lee, H.-M.; Kim, H.-S.; Moon, E.-S.; Park, J.-O.; Hwang, B.-H.; Son, J.-H.; Moon, S.-H. A validated finite element analysis of nerve root stress in degenerative lumbar scoliosis. Medical & Biological Engineering & Computing 2009, 47, 599–605. 62. Greaves, C. Y.; Gadala, M. S.; Oxland, T. R. A Three-Dimensional Finite Element Model of the Cervical Spine with Spinal Cord: An Investigation of Three Injury Mechanisms. Annals of Biomedical Engineering 2008, 36 (3), 396–405. 63. Jones, A. C.; Wilcox, R. K. Assessment of Factors Influencing Finite Element Vertebral Model Predictions. Journal of Biomechanical Engineering 2007, 129, 898–903. 64. Silva, M. J.; Keaveny, T. M.; Hayes, W. C. Load Sharing Between the Shell and Centrum in the Lumbar Vertebral Body. Spine 1997, 22, 140–150. 65. Liebschner, M. A.; Rosenberg, W. S.; Keaveny, T. M. Effects of Bone Cement Volume and Distribution on Vertebral Stiffness After Vertebroplasty. Spine 2001, 26 (14), 1547–1554.
30
66. Higgins, K. B.; Sindall, D. R.; Cuitino, A. M.; Langrana, N. A. Biomechanical Alterations in Intact Osteoporotic Spine Due to Synthetic Augmentation: Finite Element Investigation. Journal of Biomechanical Engineering 2007, 129 (4), 575–585. 67. Whyne, C. M.; Hu, S. S.; Lotz, J. C. Burst Fracture in the Metastatically Involved Spine. Spine 2003, 28 (7), 652–660. 68. Sun, K.; Liebschner, M. A. Evolution of Vertebroplasty: A Biomechanical Perspective. Annals of Biomedical Engineering 2004, 32 (1), 77–91. 69. Wijayathunga, V. N.; Jones, A.; Oakland, R. J.; Furtado, N. R.; Hall, R. M.; Wilcox, R. K. Development of specimen-specific finite element models of human vertebrae for the analysis of vertebroplasty. Proceedings of the Institution of Mechanical Engineers, Part H: Journal of Engineering in Medicine 2008, 222 (2), 221–228. 70. Buckley, J. M.; Loo, K.; Motherway, J. Comparison of quantitative computed tomography-based measures in predicting vertebral compressive strength. Bone 2007, 40 (3), 767–774. 71. Crawford, R. P.; Cann, C. E.; Keaveny, T. M. Finite element models predict in vitro vertebral body compressive strength better than quantitative computed tomography. Bone 2003, 33, 744–750. 72. Imai, K.; Ohnishi, I.; Bessho, M.; Nakamura, K. Nonlinear Finite Element Model Predicts Vertebral Bone Strength and Fracture Site. Spine 2006, 31 (16), 1789–1794. 73. Liebschner, M. A. K.; Kopperdahl, D. L.; Rosenberg, W. S.; Keaveny, T. M. Finite Element Modeling of the Human Thoracolumbar Spine. Spine 2003, 28 (6), 559–565. 74. Silva, M. J.; Keaveny, T. M.; Hayes, W. C. Computed Tomography-Based Finite Element Analysis Predicts Failure Loads and Fracture Patterns for Vertebral Sections. Journal of Orthopaedic Research 1998, 16, 300–308. 75. Meakin, J. R.; Hukins, D. Replacing the nucleus pulposus of the intervertebral disk: prediction of suitable properties of a replacement material using finite element analysis. Journal of Materials Science: Materials in Medicine 2001, 12, 207–213. 76. Mascio, V. D.; Bellini, C. M.; Galbusera, F.; Raimondi, M. T.; Brayda-Bruno, M.; Assietti, R. Lumbar total disc replacement: A numerical study. Journal of Applied Biomaterials & Biomechanics 2010, 8 (2), 97–101. 31
77. Schroeder, Y.; Wilson, W.; Huyghe, J. M.; Baaijens, F. P. T. Osmoviscoelastic finite element model of the intervertebral disc. European Spine Journal 2005, 15 (Suppl. 3), S361–S371. 78. Ferguson, S. J.; Ito, K.; Nolte, L. P. Fluid flow and convective transport of solutes within the intervertebral disc. Journal of Biomechanics 2004, 37, 213–221. 79. Espino, D. M.; Meakin, J. R.; Hukins, D. W. L.; Reid, J. E. Stochastic Finite Element Analysis of Biological Systems: Comparison of a Simple Intervertebral Disc Model with Experimental Results. Computer Methods in Biomechanics and Biomedical Engineering 2003, 6 (4), 243–248. 80. Argoubi, M.; Shirazi-Adl, A. Poroelastic Creep Response Analysis of a Lumbar Motion Segment in Compression. Journal of Biomechanics 1996, 29 (10), 1331–1339. 81. Schmidt, H.; Heuer, F.; Simon, U.; Kettler, A.; Rohlmann, A.; Claes, L.; Wilke, H.-J. Application of a new calibration method for a three-dimensional finite element model of a human lumbar annulus fibrosus. Clinical Biomechanics 2006, 21, 337–344. 82. Rohlmann, A.; Burra, N. K.; Zander, T.; Bergmann, G. Comparison of the effects of bilateral posterior dynamic and rigid fixation devices on the loads in the lumbar spine: a finite element analysis. European Spine Journal 2007, 16 (8), 1223–1231. 83. Kim, Y. Finite Element Analysis of Anterior Lumbar Interbody Fusion. Spine 2007, 32 (23), 2558–2568. 84. Fantigrossi, A.; Galbusera, F.; Raimondi, M.; Sassi, M.; Fornari, M, Biomechanical analysis of cages for posterior lumbar interbody fusion. Medical Engineering & Physics 2007, 29 (1), 101–109. 85. Zander, T.; Rohlmann, A.; Calisse, J.; Bergmann, G. Estimation of muscle forces in the lumbar spine during upper-body inclination. Clinical Biomechanics 2001, 16 (Suppl. 1), S73–S80. 86. Nataranjan, R. N.; Williams, J. R.; Lavender, S. A.; Andersson, G. B. J. Poro-elastic finite element model to predict the failure progression in a lumbar disc due to cyclic loading. Computers & Structures 2007, 85, 1142–1151. 87. Wu, J. S. S.; Chen, J. H. Clarification of the mechanical behavior of spinal motion segments through a three-dimensional poroelastic mixed finite element model. Medical Engineering & Physics 1996, 18 (3), 215–224. 32
88. Lee, C.-K.; Kim, Y. E.; Lee, C.-S.; Hong, Y.-M.; Jung, J.-M.; Goel, V. K. Impact Response of the Intervertebral Disc in a Finite-Element Model. Spine 2000, 25 (19), 2431–2439. 89. Nicolella, D. P.; Francis, W. L. Personal Communication, 2011. 90. Teo, E. C.; Ng, H. W. First cervical vertebra (atlas) fracture mechanism studies using finite element method. Journal of Biomechanics 2001, 34, 13–21. 91. Adams, M. A.; Dolan, P. Time-dependent changes in the lumbar spine’s resistance to bending. Clinical Biomechanics 1996, 11 (4), 194–200. 92. Li, X.-F.; Dai, L.-Y. Three-Dimensional Finite Element Model of the Cervical Spinal Cord. Spine 2009, 34 (11), 1140–1147. 93. Dar, F. H.; Meakin, J. R.; Aspden, R. M. Statistical methods in finite element analysis. Journal of Biomechanics 2002, 35 (9), 1155–1161. 94. Rohlmann, A.; Mann, A.; Zander, T.; Bergmann, G. Effect of an artificial disc on lumbar spine biomechanics: a probabilistic finite element study. European Spine Journal 2009, 18, 89–97. 95. Ng, H. W.; Teo, E. C. Probabilistic Design Analysis of the Influence of Material Property on the Human Cervical Spine. Journal of Spinal Disorders & Techniques 2004, 17 (2), 123–133. 96. Lee, K. K.; Teo, E. C. Material Sensitivity Study on Lumbar Motion Segment (L2 -L3 ) Under Sagittal Plane Loadings Using Probabilistic Method. Journal of Spinal Disorders & Techniques 2005, 18 (2), 163–170. 97. NASA, Structural Analysis Made “NESSUSary”. http://www.sti.nasa.gov/tto/ Spinoff2005/ct 8.html, 2005 (accessed: 15 December 2010). 98. Shender, B. S.; Paskoff, G. Overview of the NAVAIR Spinal Injury Mitigation Program. In Proceedings of the Forty Third Annual SAFE Association Symposium; Salt Lake City, UT, 2005. 99. Rohlmann, A.; Boustani, H. N.; Bergmann, G.; Zander, T. A probabilistic finite element analysis of the stresses in the augmented vertebral body after vertebroplasty. European Spine Journal 2010, 19 (9), 1585–1595.
33
100. Rohlmann, A.; Boustani, H. N.; Bergmann, G.; Zander, T. Effect of a pedicle-screw-based motion preservation system on lumbar spine biomechanics: A probabilistic finite element study with subsequent sensitivity analysis. Journal of Biomechanics 2010, 43, 2963–2969. 101. Thacker, B. H.; Francis, W. L.; Nicolella, D. P. Model Validation and Uncertainty Quantification Applied to Cervical Spine Injury Assessment. In Computational Uncertainty in Military Vehicle Design (pp. 26-1 26-30). Meeting Proceedings RTO-MP-AVT-147, Paper 26; Neuilly-sur-Seine, France: RTO, 2007. 102. Scheidler, M.; Kraft, R. Inertial effects in compression Hopkinson bar tests on soft materials. In In Proceedings of the 1st ARL Ballistic Protection Technologies Workshop; U.S. Army Research Laboratory: Aberdeen Proving Ground, MD, 2010. 103. Scheidler, M.; Fitzpatrick, J.; Kraft, R. Optimal Pulse Shapes for SHPB Tests on Soft Materials. In In Proceedings of the 2011 SEM Annual Conference and Exposition on Experimental and Applied Mechanics; Springer: New York, NY, 2011. 104. Anderson, A. E.; Ellis, B. J.; Weiss, J. A. Verification, validation and sensitivity studies in computational biomechanics. Computer Methods in Biomechanics and Biomedical Engineering 2007, 10 (3), 171–184.
34
Appendix A. Verification, Sensitivity and Validation
Models must establish credibility before clinicians and scientists can use the simulations to extrapolate information and make decisions from information and results produced by the model. Specifically, an analyst must (1) assess the numerical accuracy of the underlying mathematical equations in the computation model, (2) evaluate the sensitivity of the model results to the input parameters, and (3) validate that the model results correspond to results of a real-world scenario (54, 104). Models are often validated through one of two methods: direct or indirect validation. Direct validation involves a comparison between a model’s prediction and an experimental test result. Indirect validation compares model results to a physical case where it is not possible to determine if conditions are the same. This includes comparisons between data from the literature, clinical trial results, and historical medical data (54). Since the first spinal finite element models were developed, significant advances have been made. Through adequate considerations of model sensitivity and validation, researchers have been able to develop more accurate and efficient simulations. Generating a precise model that incorporates verification, analysis sensitivity, and validation is a difficult task, especially for the complex spinal structure. However by doing so, the gaps between clinical medicine, experimental biology, and computational biomechanics will decrease.
35
NO. OF COPIES ORGANIZATION 1 DEFENSE TECHNICAL (PDF INFORMATION CTR ONLY) DTIC OCA 8725 JOHN J KINGMAN RD STE 0944 FORT BELVOIR VA 22060-6218 1
DIRECTOR US ARMY RESEARCH LAB IMNE ALC HRR 2800 POWDER MILL RD ADELPHI MD 20783-1197
1
DIRECTOR US ARMY RESEARCH LAB RDRL CIO LL 2800 POWDER MILL RD ADELPHI MD 20783-1197
1
DIRECTOR US ARMY RESEARCH LAB RDRL CIO MT 2800 POWDER MILL RD ADELPHI MD 20783-1197
1
DIRECTOR US ARMY RESEARCH LAB RDRL D 2800 POWDER MILL RD ADELPHI MD 20783-1197
36
NO. OF COPIES ORGANIZATION
NO. OF COPIES ORGANIZATION
ABERDEEN PROVING GROUND 41
2
FT. DETRICK, MD 21702-5012 M LEGGIERI R GUPTA
DIR USARL RDRL WMP S SCHOENFELD RDRL WMP F N GNIAZDOWSKI R KARGUS R GUPTA A FRYDMAN E FIORVONTE RDRL WMP C T W BJERKE RDRL CIH C P CHUNG RDRL WML H B SCHUSTER RDRL WMP B C HOPPEL B LEAVY S R BILYK D CASEM J CLAYTON D DANDEKAR M GRINFELD Y I HUANG B LOVE M RAFTENBERG M SCHEIDLER T WEERISOORIYA D POWELL R KRAFT (10 CPS) RDRL-SL R COATES RDRL-SLB-W W MERMAGEN M TEGMEYER C KENNEDY P GILLICH A BREUER L ROACH R SPINK A KULAGA RDRL-HRS-C K MCDOWELL K OIE B LANCE W HAIRSTON DoD BLAST INJURY RESEARCH PROGRAM COORDINATING OFFICE USAMRMC, ATTN: MCMR-RTB
37
1
HUMAN SYSTEMS DEPT CODE 4656 BLDG 2187 SUITE 2280A NAVAL AIR WARFARE CENTER AIRCRAFT DIVISION 48110 Shaw Road UNIT 5 PATUXENT RIVER, MD 20670-1906 B SHENDER
1
AMC-TARDEC ATTN: RDTA-RS BUILDING 200C, ROOM 1150 WARREN, MI 48397 R SCHERER
1
ANTON E. DMITRIEV DEPARTMENT OF SURGERY, A3020 UNIFORMED SERVICES UNIVERSITY 4301 JONES BRIDGE ROAD BETHESDA, MD 20814
1
ANDRZEJ J. PRZEKWAS CFD RESEARCH CORPORATION 215 WYNN DRIVE HUNTSVILLE, ALABAMA 35805
1
RABIH TANNOUS BAE SYSTEMS 9113 LE SAINT DRIVE FAIRFIELD, OH 45014-5453
1
PHILIP DUDT NAVEL SURFACE WARFARE CENTER, CODE 664 9500 MACARTHUR DRIVE WEST BETHESDA, MD 20817
1
MICHAEL CODEGA AMC-NSRDEC KANSAS STREET NATICK, MA 01760
2
SOUTHWEST RESEARCH INSTITUTE MECHANICAL AND MATERIALS ENGINEERING DIVISION MATERIALS ENGINEERING DEPARTMENT 6220 CULEBRA ROAD SAN ANTONIO, TX 78238 D NICOLELLA W FRANCIS
NO. OF COPIES ORGANIZATION 3
1
1
1
1
5
1
NO. OF COPIES ORGANIZATION
JTAPIC PROGRAM OFFICE U.S. ARMY MEDICAL RESEARCH AND MATERIAL COMMAND ATTN: MRMC-RTB 504 SCOTT STREET FORT DETRICK, MD 21702-5012 J USCILOWICZ W LEI F LEBEDA
CAMBRIDGE, MASSACHUSETTS 02139-4307
KEN L. MONSON THE UNIVERSITY OF UTAH 50 S. CENTRAL CAMPUS DRIVE 2132 MERRILL ENGINEERING BLDG SALT LAKE CITY, UTAH 84112 NAMAS CHANDRA UNIVERSITY OF NEBRASKA 114G OTHMER HALL P.O. BOX 880642 LINCOLN, NE 68588-0642 SHANE SCHUMACHER SANDIA NATIONAL LABORATORIES NANOSCALE AND REACTIVE PROCESSES P.O. BOX 5800, MS 0836 ALBUQUERQUE, NEW MEXICO 87185-0836 ADAM FOURNIER U.S. ARMY ABERDEEN TEST CENTER ATTN: TEDT-AT-SLB 400 COLLERAN ROAD ABERDEEN PROVING GROUND, MARYLAND 21005-5059 THE JOHNS HOPKINS UNIVERSITY APPLIED PHYSICS LABORATORY 11100 JOHNS HOPKINS ROAD LAUREL, MARYLAND 20723-6099 ANDREW MERKLE MORGANA TREXLER ANDY LENNON JACK ROBERTS TIM HARRIGAN RAUL A. RADOVITZKY AERONAUTICS AND ASTRONAUTICS MASSACHUSETTS INSTITUTE OF TECHNOLOGY
38
1
PAUL E. RAPP DIRECTOR, TRAUMATIC INJURY RESEARCH PROGRAM DEPARTMENT OF MILITARY AND EMERGENCY MEDICINE UNIFORMED SERVICES UNIVERSITY OF THE HEALTH SCIENCES 4301 JONES BRIDGE ROAD BETHESDA, MARYLAND 20814-4799
1
JOHN M. GETZ U.S. ARMED FORCES MEDICAL EXAMINER SYSTEM 1413 RESEARCH BLVD. ROCKVILLE, MARYLAND 20850
1
ALAN HEPPER DSTL BIOMEDICAL SCIENCES RM 1A, BLDG. 245 PORTON DOWN SALISBURY, WILTSHIRE SP4 OJQ UNITED KINGDOM
3
DRDC VALCARTIER 2459, PIE-XI BLVD. NORTH QUEBEC, QC G3J 1X5 CANADA KEVIN WILLIAMS AMAL BOUAMOUL LUCIE MARTINEAU
2
CENTER FOR INJURY BIOMECHANICS WAKE FOREST UNIVERSITY MEDICAL CENTER BLVD. WINSTON-SALAM, NC 27157 JOEL STITZEL F. SCOTT GAYZIK
1
DANIEL WISE HENRY JACKSON FOUNDATION U.S. ARMY AEROMEDICAL RESEARCH LABORATORY 6901 ANDREWS AVENUE FORT RUCKER, AL 36362-0577
1
TOM RADTKE HUMAN PROTECTION AND PERFORMANCE DIVISION DEFENCE SCIENCE AND TECHNOLOGY ORGANISATION
NO. OF COPIES ORGANIZATION
NO. OF COPIES ORGANIZATION
DEPARTMENT OF DEFENCE BLDG 109, 506 LORIMER STREET FISHERMANS BEND, VICTORIA 3207 AUSTRALIA 2
DEPARTMENT OF MECHANICAL ENGINEERING THE JOHNS HOPKINS UNIVERSITY LATROBE 122 3400 NORTH CHARLES STREET BALTIMORE, MD 21218 K.T. RAMESH VICKY NGUYEN
39
I NTENTIONALLY LEFT BLANK .
40
